Frequent GEODYN Problems
---------------------------
1) Light Time Problems in Interplanetary Runs:
1a) RANGE AND/OR DOPPLER OBSERVATIONS NEAR THE INITIAL EPOCH:
It is important to remember that DSN Doppler and range measurements
are time tagged at the epoch when the signal is received at the tracking
station. When any measurement is processed, the epoch(s) at which the
signal are at the satellite must be modeled using satellite positions
and/or velocities. The satellite epoch will be a light time earlier than
the DSN Doppler or range time tag. The satellite positions and velocities
are computed by numerical integration and are only available epochs after
the initial epoch specified on the EPOCH card. Range and doppler observations
that are within a light time of the initial epoch should be deleted.
1b) MIXING RANGE AND/OR DOPPLER OBSERVATIONS WITH ALTIMETRY OR CAMERA DATA IN
INTERPLANETARY RUNS: To avoid problems limit the maximum number of observations
in a block (columns 11-14 of the VECOPT card) and maximize the numerical integrator's
window of interpolation (columns 15-17 of the VECOPT card). The detailed explanation
follows: GEODYN (TDF and IIS) will order observations in order of their nominal time
tag. It is important to remember that range and Doppler observations are time tagged
at the epoch that the tracking station receives the signal while camera and altimeter
ranges are time tagged at satellite time. As a result, range and Doppler can have
large light time issues while altimeter and camera observations do not. When Doppler
and range observations are mixed with satellite time tagged observations in interplanetary
runs, light time issues force GEODYN to model satellite epochs out of chronological
order. The problem is compounded by the fact that GEODYN TDF & IIS group observations
into blocks. GEODYN will not integrate backwards to accommodate satellite times that
are not in chronological order. Even so, problems can be avoided by using settings
on the VECOPT card. GEODYN can successfully process satellite epochs that are not in
chronological order, but only if the window of interpolation associated with the
integrator is large enough to span events that are out of order. The window
(time interval) of interpolation is the number of back values carried by the
integrator (columns 15-17 of the VECOPT card) multiplied by the integration step size.
The worst case of out of order observations will occur if a block of Doppler
observations follows a large block of altimeter or camera observations. First the
integrator will have been advanced to the time tag of the final altimeter or camera
observation in the block. After the altimeter or camera observation have been
processed, the block of Doppler observations will be processed. The earliest satellite
event will be for the first Doppler observation in the block at the beginning of the
counting interval. This will occur at a time earlier than the observation time tag by
the counting interval plus the light time. This earliest satellite event time for the
Doppler is likely to be significantly earlier than the time tag of the final altimeter
of camera observation of the previous block. This interval can be particularly long
because GEODYN TDF and IIS tend to group long passes of altimeter observations together
into large blocks. The problem can be mitigated by using the minimum value (11) of the
maximum number of observations in a block (columns 11-14 of the VECOPT card). In addition
to limiting the number of observations in a block, the window of interpolation should be
maximized (columns 15-17 of the VECOPT card) to cover the light time plus some extra.